| 18 Jun 2024 |
@cnx:loang.net | developers have been spoiled with really powerful amd64 compooters, laptops from almost a decade ago are still really usable today and they're much more powerful than the mid-level risc-v soc we've seen so far | 17:37:17 |
Pratham Patel (you can mention me) | In reply to @cnx:loang.net the jh7110 is very well mainlined, but true it's quite weak for a laptop not that it's weak for a laptop but that it's "profile" (if you can even call it that) is RV64GC which is very basic | 17:44:01 |
Pratham Patel (you can mention me) | while the "real" profiles which will be on servers (yes they're coming soon) will be using RVA23 | 17:44:30 |
Tristan Ross | What's the difference between RVA23 and RV64GC? | 17:46:35 |
Pratham Patel (you can mention me) | Which has all the fancy extensions and more importantly, platform features like Sv57 | 17:47:06 |
Pratham Patel (you can mention me) | In reply to @rosscomputerguy:matrix.org What's the difference between RVA23 and RV64GC? https://github.com/riscv/riscv-profiles/blob/main/src/rva23-profile.adoc | 17:47:12 |
| * Pratham Patel (you can mention me) logs off for the night (IST) | 17:47:40 |
Tristan Ross | In reply to @thefossguy:matrix.org https://github.com/riscv/riscv-profiles/blob/main/src/rva23-profile.adoc Oh nice | 17:47:54 |
Tristan Ross | Floating points and vectors will be great to have | 17:48:23 |
Tristan Ross | Hopefully chips go with a good implementation of things in the silicon | 17:48:56 |
Pratham Patel (you can mention me) | before I log off; think of RV64GC as Arm-v8 and RVA23 as SBBR or whatever they call their Server Platform base thingy | 17:49:18 |
Pratham Patel (you can mention me) | In reply to @rosscomputerguy:matrix.org Floating points and vectors will be great to have g already contains imafd ;) | 17:49:46 |
Tristan Ross | In reply to @thefossguy:matrix.org before I log off; think of RV64GC as Arm-v8 and RVA23 as SBBR or whatever they call their Server Platform base thingy Ooh, hopefully this brings RISC-V chips a bit closer to being more performant | 17:51:22 |
Tristan Ross | In reply to @thefossguy:matrix.org g already contains imafd ;) Didn't realize that | 17:51:31 |
Pratham Patel (you can mention me) | ranted here in more detail
https://matrix.to/#/!GhnNOKFLyvWBmnjqRS:fedoraproject.org/$9FGD8-67N7zQIhfMbWZKBqCa3qT9zq6l25iGim11ifY?via=fedoraproject.org&via=matrix.org&via=fedora.im | 17:53:20 |
Alex | In reply to @rosscomputerguy:matrix.org Didn't realize that More precisely, G is shorthand for IMAFD_Zicsr_Zifencei. | 18:21:37 |
Tristan Ross | Oh, I know zicsr and zifencei | 18:35:38 |
| 22 Jun 2024 |
| nyanbinary 🏳️⚧️ joined the room. | 01:21:39 |
| throwaway account joined the room. | 05:43:33 |
| 24 Jun 2024 |
Mic92 | We are just one pull request away from having our first image for nixos-anywhere: https://github.com/NixOS/nixpkgs/pull/321991 | 13:27:36 |
Mic92 | However I currently don't have hardware that I can kexec on, so I would appreciate if someone could run a test of it. | 13:28:08 |
Mic92 | Image itself will be here: https://github.com/nix-community/nixos-images/pull/243 | 13:28:19 |
Pratham Patel (you can mention me) | Oh wow, didn't know we came this far! | 13:29:14 |
Mic92 | nix-community.cachix.org will than also cache some core-components for cross compiling NixOS from x86 to riscv64. | 13:29:20 |
Pratham Patel (you can mention me) | *we've come | 13:29:20 |
[0x4A6F] | Currently not near my devices. But it is build and I will check soonish. | 14:04:18 |
| Dr. Worm changed their profile picture. | 23:17:28 |
| 25 Jun 2024 |
Tristan Ross | In reply to @joerg:thalheim.io nix-community.cachix.org will than also cache some core-components for cross compiling NixOS from x86 to riscv64. What about cross compiling from aarch64 to riscv64? | 04:20:53 |
@cnx:loang.net | why x86 and aarch64 when amd64 chips with good single thread perf are much more common? | 09:34:12 |
Alex | In reply to @cnx:loang.net why x86 and aarch64 when amd64 chips with good single thread perf are much more common? I'm pretty sure amd64 is what Mic was referring to when they said "x86". | 10:48:28 |